Storing vibration data in time series


From what I can see your time stamp is down to a millisecond.
If I wanted to store some time series data that is sampled at 5kHz (0.2ms), do I have any way to do that?


We typically recommend that high-frequency data, like vibration, be processed at the edge. The edge device can then report higher-level insights to the cloud based on the analysis.

The reason for this is a simple matter of data volume. At 0.2ms or 5,000 points per second, that’s 432MM data points per day. If each data point is only one integer, that 4 bytes per point, or ~1.6GB per day of data being generated.

Our underlying time-series database does not support resolutions greater than 1ms, so directly storing sub-millisecond data is not possible.


Had a feeling that would be your answer.